back to article Can't wait for Nvidia? Try these Italian baby ARM clusters with GPU options

Nvidia has made no secret about wanting to be a player in the supercomputer racket both on the GPU and CPU sides of a hybrid system. The company launched "Project Denver" nearly two years ago to create a Nvidia-branded chip, which will see Denver ARM processors timed to market with the future "Maxwell" GPUs two years from now. …

COMMENTS

This topic is closed for new posts.
  1. Anonymous Coward
    Anonymous Coward

    ..not another box

    nice idea. now just provide the kit on PCIX/PCIe boards so we can plug them into our chosen servers and use the oomph for processing when we need.

  2. Anonymous Coward
    Anonymous Coward

    RE: The ARM boards that are in the E4 Computer servers are the same ones that are used in the experimental "Mont Blanc" machine, which pairs quad-core Tegra3 ARM-based chips with Nvidia modile GPUs, at the Barcelona Supercomputing Center in Spain.

    Not actually true - the MontBlanc machine will use the Samsung Exynos 5 Dual chips - 2 x Cortex-A15 + ARM's OpenCL enabled Mali T604 GPU :

    http://www.bsc.es/about-bsc/press/bsc-in-the-media/mont-blanc-project-selects-samsung-exynos-5-processor

    1. tpm (Written by Reg staff)

      I did not know that. Tweaked the story to reflect the change. Thanks.

  3. Anonymous Coward
    Anonymous Coward

    Not sure about this...

    supposed to pair the Qseven card from SECO with a single Tegra3 processor, which has four Cortex-A9 cores running at 1.5GHz plus a fifth baby core for management. (This is the so-called "big.Little" architecture that ARM is espousing to support different-sized workloads with a single chip.)

    big.LITTLE uses different core designs for low power vs. high performance. All 5 cores in the Tegra3 are A9s. The 5th core uses a different fab process that makes it more efficient but also prevents it from running the same clocks as the other 4 cores.

    big.LITTLE either pairs A15 and A7 cores, or eventually A57 and A53. AFAIK there are no big.LITTLE procs commonly available. Tegra3's companion core is similar, in intent at least, but there is no "big.LITTLE" for A9s.

  4. hplasm
    Coat

    Baby ARM

    Holding an Apple?

  5. Timbo
    Coat

    "But pairing an ARM processor with a GPU – essentially a modern-style, outboard math coprocessor like Intel used to offer in a special socket for x86 CPUs before they were brought on-chip with the 80486SX and Pentium chips......"

    Errr.....not quite true - the 486DX has the FPU built-in.

    For the 486SX, you had to buy the the 487SX, to get an FPU, (though by plugging in the 487, it then disabled the onboard 486SX.... :( )

    1. Aitor 1

      damn, you were faster than me!!

      Yes 486SX processors had the FPU disconnected... you needed the 487.

      I had a nice 486DX66, by the way...

  6. kukreknecmi

    The Article is jumping from one branch to another, making it too hard to follow the scattered information for thoose who are a bit unfamiliar of to theese kind of stuff.

This topic is closed for new posts.

Other stories you might like